I hadn't seen that, thanks for it.
I'm working for an alternative energy startup
in India, and we're trying to make a modular CSP system using heliostats and the Arduino platform. The Indian monsoon season throws off the cost-efficiency (and reliability) a bit.
Your heliostat program is quite amazing, but being limited to 16 units does make it a little small for what we'd like to try, while being an excellent proof of concept. I.e. something like a few square meters of mirrors - a larger number of small ones, perhaps each with its own motor, to allow more concentration. The main applications here would be heat, not power generation e.g. steam for a spa. As you can imagine, a lot of electricity goes to use for that. We're still working out the part costs.
We haven't made a model yet, but pure geometry says the LDR-guided heliostats should work, at least in my mind. With the LDRs and mirror normals aiming directly at the target, set that to be the "reset" position in all axes - then set it loose. The mirror normals should always bisect the angle between the line mirror->Sun (dictated by the LDR) and the line mirror->target. The mechanical solution to that should be a simple 2:1 gear reduction between the LDRs and mirror movement in both axes. I could be wrong, but I don't see how.